************************************************************************************************ GROUP COMPARISON AT BASELINE AND FOLLOW-UP WITH BOOTSTRAP
**** This file generates output tables for group comparison at baseline with bootstrap 
**** standard errors.
**** Version: 20/05/2014********************************************************************************************	
			
************************************************************************************************ HOUSEHOLD LEVEL VARIABLES
********************************************************************************************

		* EXPENDITURES
		
			* Create table for mean differences between treatment status for different data collection waves
			
				treattest_bs $Expenditure_shares_var
							
				esttab  using "Outputs/Expenditure shares at baseline and follow-up (with bootstrap).tex", f replace compress noobs nonumb width(1\textwidth) ///
				cells("m_mu_1_b(fmt(3)) m_mu_2_b(fmt(3)) m_coef_c_b_bs(fmt(3) star pvalue(m_d_p_c_b_bs)) m_mu_1_f(fmt(3)) m_mu_2_f(fmt(3)) m_coef_c_f_bs(fmt(3) star pvalue(m_d_p_c_f_bs))" ///
				"m_sd_1_b(fmt(3) par([ ])) m_sd_2_b(fmt(3) par([ ])) m_d_se_c_b_bs(fmt(3) par) m_sd_1_f(fmt(3) par([ ])) m_sd_2_f(fmt(3) par([ ])) m_d_se_c_f_bs(fmt(3) par)") label star(* 0.1 ** 0.05 *** 0.01) ///
				stats(N1 N2 N3 N4 N5 N6 N7 N8, layout("@ @ @ @ @ @ @ @") labels("Observations") fmt(0)) ///
				mgroups("\multicolumn{3}{c}{\textit{Baseline}} & \multicolumn{3}{c}{\textit{Follow-up}}",  pattern(1 0 0 1 0 0) span) ///
				collabels("HH" "Mother" "Diff" "HH" "Mother" "Diff") ///
				ti("Expenditure and budget shares by treatment status (bootstrap standard errors)\label{ExpAllBs}") ///
				addn("\begin{minipage}[t]{0.9\textwidth}{\smaller Standard deviations in brackets, standard errors in parenthesis. \sym{***} denotes significance at 1\%, \sym{**} at 5\%, and \sym{*} at 10\%. The standard errors on the differences are estimated from running the corresponding least squares regression on treatment status allowing for the errors to be clustered at municipality level. Treatment status is equal to 1 if the transfer is made to mothers and zero otherwise. Bootstrap is performed with clusters at municipality level and with 500 repetitions.} \end{minipage}")
	
		* CONSUMPTION
			
			* Create table for mean differences between treatment status for different data collection waves
			
				treattest_bs $Cons_shares_var
			
				esttab  using "Outputs/Consumption shares at baseline and follow-up (with bootstrap).tex", f replace compress noobs nonumb width(1\textwidth) ///
				cells("m_mu_1_b(fmt(3)) m_mu_2_b(fmt(3)) m_coef_c_b_bs(fmt(3) star pvalue(m_d_p_c_b_bs)) m_mu_1_f(fmt(3)) m_mu_2_f(fmt(3)) m_coef_c_f_bs(fmt(3) star pvalue(m_d_p_c_f_bs))" ///
				"m_sd_1_b(fmt(3) par([ ])) m_sd_2_b(fmt(3) par([ ])) m_d_se_c_b_bs(fmt(3) par) m_sd_1_f(fmt(3) par([ ])) m_sd_2_f(fmt(3) par([ ])) m_d_se_c_f_bs(fmt(3) par)") label star(* 0.1 ** 0.05 *** 0.01) ///
				stats(N1 N2 N3 N4 N5 N6 N7 N8, layout("@ @ @ @ @ @ @ @") labels("Observations") fmt(0)) ///
				mgroups("\multicolumn{3}{c}{\textit{Baseline}} & \multicolumn{3}{c}{\textit{Follow-up}}",  pattern(1 0 0 1 0 0) span) ///
				collabels("HH" "Mother" "Diff" "HH" "Mother" "Diff") ///
				ti("Consumption and consumption shares by treatment status (bootstrap standard errors)\label{ConsAllBs}") ///
				addn("\begin{minipage}[t]{0.9\textwidth}{\smaller Standard deviations in brackets, standard errors in parenthesis. \sym{***} denotes significance at 1\%, \sym{**} at 5\%, and \sym{*} at 10\%. The standard errors on the differences are estimated from running the corresponding least squares regression on treatment status allowing for the errors to be clustered at municipality level. Treatment status is equal to 1 if the transfer is made to mothers and zero otherwise. Bootstrap is performed with clusters at municipality level and with 500 repetitions.} \end{minipage}")
